Evaluation of photoshop based image analysis in cytologic diagnosis of pleural fluid in comparison with conventional modalities

摘要

Background The aim of this study is to investigate and compare the results of digital image analysis in pleural effusion cytology samples with conventional modalities. Materials and Methods In this cross‐sectional study, 53 pleural fluid cytology smears from Qaem hospital pathology department, located in Mashhad, Iran were investigated. Prior to digital analysis, all specimens were evaluated by two pathologists and categorized into three groups as: benign, suspicious, and malignant. Using an Olympus microscope and Olympus DP3 digital camera, digital images from cytology slides were captured. Appropriate images ( n ?=?130) were separately imported to Adobe Photoshop CS5 and parameters including area and perimeter, circularity, Gray Value mean, integrated density, and nucleus to cytoplasm area ratio were analyzed. Results Gray Value mean, nucleus to cytoplasm area ratio, and circularity showed the best sensitivity and specificity rates as well as significant differences between all groups. Also, nucleus area and perimeter showed a significant relation between suspicious and malignant groups with benign group. Whereas, there was no such difference between suspicious and malignant groups. Conclusion We concluded that digital image analysis is welcomed in the field of research on pleural fluid smears as it can provide quantitative data to apply various comparisons and reduce interobserver variation which could assist pathologists to achieve a more accurate diagnosis.